Astrophys Space Sci (2014) 352:341–352 DOI 10.1007/s10509-014-1961-4 ORIGINAL ARTICLE ACRIM3 and the Total Solar Irradiance database Richard C. Willson Received: 16 December 2013 / Accepted: 14 February 2014 / Published online: 16 May 2014 © The Author(s) 2014. This article is published with open access at Springerlink.com Abstract The effects of scattering and diffraction on the and its +0.037 %/decade TSI trend during solar cycles 21– observations of the ACRIMSAT/ACRIM3 satellite TSI 23 is the most likely correct representation of the extant monitoring mission have been characterized by the preflight satellite TSI database. The occurrence of this trend during calibration approach for satellite total solar irradiance (TSI) the last decades of the 20th century supports a more robust sensors implemented at the LASP/TRF (Laboratory for At- contribution of TSI variation to detected global temperature mospheric and Space Physics/Total Solar Irradiance Ra- increase during this period than predicted by current climate diometer Facility). The TRF also calibrates the SI (Interna- models. tional System of units) traceability to the NIST (National In- Keywords Total Solar Irradiance · SI Calibration · Decadal stitute of Standards and Technology) cryo-radiometric scale. Trends ACRIM3’s self-calibration agrees with NIST to within the uncertainty of the test procedure (∼500 ppm). A correc- ∼ tion of 5000 ppm was found for scattering and diffraction 1 Introduction that has significantly reduced the scale difference between the results of the ACRIMSAT/ACRIM3 and SORCE/TIM The Earth’s climate regime is determined by the total solar satellite experiments. Algorithm updates reflecting more irradiance (TSI) and its interactions with the Earth’s atmo- than 10 years of mission experience have been made that sphere, oceans and landmasses. Evidence from 35 years of further improve the ACRIM3 results by eliminating some satellite TSI monitoring and solar activity data has estab- thermally driven signal and increasing the signal to noise lished a paradigm of direct relationship between TSI and ratio. The result of these changes is a more precise and de- solar magnetic activity. (Willson et al. 1981; Willson and tailed picture of TSI variability. Comparison of the results Hudson 1991; Willson 1997, 1984; Frohlich and Lean 1998; from the ACRIM3, SORCE/TIM and SOHO/VIRGO satel- Scafetta and Willson 2009; Kopp and Lean 2011a, 2011b) lite experiments demonstrate the near identical detection of This paradigm, together with the satellite record of TSI and TSI variability on all sub-annual temporal and amplitude proxies of historical climate and solar variability, support the scales during the TIM mission. The largest occurs at the connection between variations of TSI and the Earth’s cli- rotational period of the primary solar activity longitudes. mate. Recent developments in preflight calibrations of TSI satellite monitors can provide significant improvements in On the decadal timescale, while ACRIM3 and VIRGO re- accuracy and traceability of the database and our under- sults exhibit close agreement throughout, TIM exhibits a standing of solar variability and its importance as a forcing consistent 500 ppm upward trend relative to ACRIM3 and of climate change. VIRGO. A solar magnetic activity area proxy for TSI has The ACRIMSAT/ACRIM3 and SORCE/TIM flight in- been used to demonstrate that the ACRIM TSI composite struments were characterized in the LASP/TRF by instru- ment proxies1 after their launches (Kopp et al. 2012). The R.C. Willson (B) 1 ACRIM, 12 Bahama Bend, Coronado, CA 92118, USA ACRIM3 and TIM backup instruments were tested and assumed to e-mail: [email protected] represent the flight versions. 342 Astrophys Space Sci (2014) 352:341–352 Fig. 1 Satellite TSI experiment results contributing to the TSI database. The inclusion of pre-1992 results in a TSI composite time series requires ‘bridging the ACRIM Gap’ with either the Nimbus7/ERB or ERBS/ERBE results. The choices produce significantly different decadal trending in the resulting TSI composite principal finding for ACRIM3 was that instrumental scatter- the ACRIM Gap through comparisons with ACRIM1 and ing and diffraction has a significant effect on its results. Af- ACRIM2 give significantly different results for TSI decadal ter reprocessing data to conform its results to this TRF test trending. The ACRIM and PMOD TSI compostite time se- result the ACRIM3 TSI values agree more closely with those ries use the ERB and ERBE results, respectively, to bridge of TIM. Other TSI flight instrumentation, including proxies the Gap. Decadal trending during solar cycles 21–23 is sig- for the SOHO/VIRGO and SOHO/DIARAD (Frohlich et al. nificant for the ACRIM composite but not for the PMOD. 1997) experiments as well as the PICARD/PREMOS flight A new TSI-specific TSI proxy database has been compiled instrument (Schmutz et al. 2013), have been characterized at that appears to resolve the issue in favor of the ACRIM com- the TRF. Application of the characterizations have resulted posite and trend. The resolution of this issue is important for in flight observations in closer agreement with TIM TSI val- application of the TSI database in research of climate change ues as was found with ACRIM3. Preflight calibrations such and solar physics. as those made at the LASP/TRF can normalize TSI satellite monitoring experiments with greater precision than sensor 1.1 The satellite TSI database self-calibrations and improve the long term traceability of the TSI database. The suite of extant TSI monitoring results is shown in One of the most perplexing issues in the 35 year satel- Fig. 1. The first satellite TSI experiment was the NOAA lite TSI database is the disagreement among TSI composite Nimbus 7/Earth Radiation Budget (Nimbus7/ERB), an ex- time series in decadal trending. Two such constructions fre- periment designed to provide data for earth radiation bud- quently cited are the ACRIM and PMOD composites. A hia- get (ERB) studies (Hickey et al. 1988;Hoytetal.1992). tus of dedicated, precision TSI monitoring was caused by the This was followed by the Solar Maximum Mission/Active termination of the SMM/ACRIM1 experiment in mid-1989 Cavity Radiometer Irradiance Monitor 1 (SMM/ACRIM1: and the delay in launching UARS/ACRIM2 until late 1991. 1980–1989), the first experiment specifically designed to During the resulting ‘ACRIM Gap’ two Earth radiation monitor TSI for the long term climate database and capa- budget experiments, the Nimbus7/ERB and ERBS/ERBE, ble of self-calibrating the on-orbit degradation of its sen- provided TSI results. The ERB and ERBE results lacked sors (Willson et al. 1981). Next was the NOAA Earth Ra- the accuracy, precision, measurement cadence and on-orbit diation Budget Satellite/Earth Radiation Budget Experiment degradation self-calibration of dedicated TSI monitors. Ad- (ERBS/ERBE: 1984–2003) that used a mission-adapted ver- ditionally, the choice of ERB or ERBS results to bridge sion of the ACRIM2 sensor design (Lee et al. 1995). Astrophys Space Sci (2014) 352:341–352 343 Fig. 2 ACRIM3 mission sensor degradation self-calibration results. The plot shows the ratios of C/A and C/B plotted relative to their initial as-launched values in units of parts per million (ppm). The shutter of sensor A operates continuously while the shutters of sensors B and C are operated brieflyevery90days.TheC/A ratio calibrates sensor A solar exposure degradation and the C/B ratio facilitates calibration of non-solar exposure degradation The Upper Atmospheric Research Satellite/Active Cav- sensors by TSI with that of electrical power in a shuttered, ity Radiometer Irradiance Monitor 2 (UARS/ACRIM2: differential mode. Interpretation of measurements in the in- 1991–2001) (Willson 1997) was intended to overlap SMM/ ternational system of units (SI) is accomplished through ACRIM1 but the Challenger disaster delayed its shuttle- electrical power calibrations and metrology of the sensor’s based launch for two years. The resulting ‘ACRIM Gap’ physical properties including its aperture area and efficiency has provided one of the major controversies in constructing for absorbing TSI (Willson 1979, 1999). TSI results are re- a composite TSI time series. This is because the ERB and ported on the ‘native scales’ of their sensors in SI units de- ERBE experiments produce significantly different results fined by the above self-calibrations and modeled diffraction when used to bridge the Gap through overlapping compar- and scattering. This type of calibration is frequently referred isons with ACRIM1 and ACRIM2 (see Sect. 3 below). to as ‘self-calibrated’ or ‘component calibration’. The Solar and Heliospheric Observatory/Variability The precision of TSI monitors is orders of magnitude less of solar IRradiance and Gravity Oscillations experiment uncertain than their accuracies in SI units so comparison (SOHO/VIRGO: 1996→) was the first of four currently on- of the overlapping results of a series of satellite monitors going TSI experiments to be deployed (Frohlich et al. 1997). can provide the maximum possible traceability for compos- The second was the ACRIM SAT/Active Cavity Radiometer ite TSI time series. The uncertainty of ACRIM3 degradation Irradiance Monitor 3 experiment (ACRIMSAT/ACRIM3: self-calibration on orbit is less than 5 ppm2/year, for exam- 2000→) (Willson 2001; Willson and Mordvinov 2003). ple (as seen in Fig. 2), while its ‘self-calibrated’ SI uncer- The third was the Solar Radiation and Climate Experi- tainty or accuracy is ∼±500 ppm (see Fig. 3). ment/Total Irradiance Monitor experiment (SORCE/TIM: → 2003 ) (Kopp et al. 2005a, 2005b). The fourth is the Total 1.3 Cryogenic radiometry Solar Irradiance Calibration Transfer Experiment (TCTE) launched in late 2013 on the US Air Force Space Test Pro- Radiometers operating at the temperature of liquid He- gram spacecraft known as STPSat-3. lium (LHe) have significantly reduced sources of thermally driven measurement uncertainties and can make TSI scale 1.2 Self-calibrating TSI sensors observations with high SI traceability.